scholarly journals Investigating the Impact of Missing Data Handling Methods on the Detection of Differential Item Functioning

2017 ◽  
pp. 1-14
Author(s):  
Hüseyin SELVİ ◽  
Devrim ÖZDEMİR ALICI
Keyword(s):  
Missing Data ◽  
Differential Item Functioning ◽  
Data Handling ◽  
Item Functioning ◽  
Handling Methods ◽  
The Impact

The Impact of Different Missing Data Handling Methods on DINA Model

2018 ◽  
Vol 7 (1) ◽  
pp. 77
Author(s):  
Seçil Ömür Sünbül
Keyword(s):  
Missing Data ◽  
Expectation Maximization Algorithm ◽  
Simulated Data ◽  
Missing At Random ◽  
Data Handling ◽  
Parameter Estimations ◽  
Dina Model ◽  
Mean Imputation ◽  
Handling Methods ◽  
The Impact

<p>In this study, it was aimed to investigate the impact of different missing data handling methods on DINA model parameter estimation and classification accuracy. In the study, simulated data were used and the data were generated by manipulating the number of items and sample size. In the generated data, two different missing data mechanisms (missing completely at random and missing at random) were created according to three different amounts of missing data. The generated missing data was completed by using methods of treating missing data as incorrect, person mean imputation, two-way imputation, and expectation-maximization algorithm imputation. As a result, it was observed that both s and g parameter estimations and classification accuracies were effected from, missing data rates, missing data handling methods and missing data mechanisms.</p>

The Impact of Different Missing Data Handling Methods on DINA Model

2018 ◽  
Vol 7 (1) ◽  
pp. 77 ◽  
Author(s):  
Seçil Ömür Sünbül
Keyword(s):  
Missing Data ◽  
Expectation Maximization Algorithm ◽  
Simulated Data ◽  
Missing At Random ◽  
Data Handling ◽  
Parameter Estimations ◽  
Dina Model ◽  
Mean Imputation ◽  
Handling Methods ◽  
The Impact

<p>In this study, it was aimed to investigate the impact of different missing data<br />handling methods on DINA model parameter estimation and classification<br />accuracy. In the study, simulated data were used and the data were generated<br />by manipulating the number of items and sample size. In the generated data,<br />two different missing data mechanisms (missing completely at random and<br />missing at random) were created according to three different amounts of<br />missing data. The generated missing data was completed by using methods<br />of treating missing data as incorrect, person mean imputation, two-way<br />imputation, and expectation-maximization algorithm imputation. As a result,<br />it was observed that both s and g parameter estimations and classification<br />accuracies were effected from, missing data rates, missing data handling<br />methods and missing data mechanisms.</p>

Missing Data Handling Methods

2018 ◽  
pp. 139-185 ◽  
Author(s):  
Craig K. Enders ◽  
Amanda N. Baraldi
Keyword(s):  
Missing Data ◽  
Data Handling ◽  
Handling Methods

scholarly journals Differential item functioning of the Boston Naming Test in cognitively normal African American and Caucasian older adults

2009 ◽  
Vol 15 (5) ◽  
pp. 758-768 ◽  
Author(s):  
OTTO PEDRAZA ◽  
NEILL R. GRAFF-RADFORD ◽  
GLENN E. SMITH ◽  
ROBERT J. IVNIK ◽  
FLOYD B. WILLIS ◽  
...  
Keyword(s):  
African American ◽  
Item Response Theory ◽  
Differential Item Functioning ◽  
Item Response ◽  
Group Performance ◽  
Response Theory ◽  
Boston Naming Test ◽  
Item Functioning ◽  
The Impact ◽  
Naming Test

AbstractScores on the Boston Naming Test (BNT) are frequently lower for African American when compared with Caucasian adults. Although demographically based norms can mitigate the impact of this discrepancy on the likelihood of erroneous diagnostic impressions, a growing consensus suggests that group norms do not sufficiently address or advance our understanding of the underlying psychometric and sociocultural factors that lead to between-group score discrepancies. Using item response theory and methods to detect differential item functioning (DIF), the current investigation moves beyond comparisons of the summed total score to examine whether the conditional probability of responding correctly to individual BNT items differs between African American and Caucasian adults. Participants included 670 adults age 52 and older who took part in Mayo’s Older Americans and Older African Americans Normative Studies. Under a two-parameter logistic item response theory framework and after correction for the false discovery rate, 12 items where shown to demonstrate DIF. Of these 12 items, 6 (“dominoes,” “escalator,” “muzzle,” “latch,” “tripod,” and “palette”) were also identified in additional analyses using hierarchical logistic regression models and represent the strongest evidence for race/ethnicity-based DIF. These findings afford a finer characterization of the psychometric properties of the BNT and expand our understanding of between-group performance. (JINS, 2009, 15, 758–768.)

Geriatric Anxiety Scale: item response theory analysis, differential item functioning, and creation of a ten-item short form (GAS-10)

2014 ◽  
Vol 27 (7) ◽  
pp. 1099-1111 ◽  
Author(s):  
Anne E. Mueller ◽  
Daniel L. Segal ◽  
Brandon Gavett ◽  
Meghan A. Marty ◽  
Brian Yochim ◽  
...  
Keyword(s):  
Older Adults ◽  
Item Response Theory ◽  
Psychometric Properties ◽  
Differential Item Functioning ◽  
Item Response ◽  
Short Form ◽  
Anxiety Scale ◽  
Response Theory ◽  
Item Functioning ◽  
The Impact

ABSTRACTBackground:The Geriatric Anxiety Scale (GAS; Segal et al. (Segal, D. L., June, A., Payne, M., Coolidge, F. L. and Yochim, B. (2010). Journal of Anxiety Disorders, 24, 709–714. doi:10.1016/j.janxdis.2010.05.002) is a self-report measure of anxiety that was designed to address unique issues associated with anxiety assessment in older adults. This study is the first to use item response theory (IRT) to examine the psychometric properties of a measure of anxiety in older adults.Method:A large sample of older adults (n = 581; mean age = 72.32 years, SD = 7.64 years, range = 60 to 96 years; 64% women; 88% European American) completed the GAS. IRT properties were examined. The presence of differential item functioning (DIF) or measurement bias by age and sex was assessed, and a ten-item short form of the GAS (called the GAS-10) was created.Results:All GAS items had discrimination parameters of 1.07 or greater. Items from the somatic subscale tended to have lower discrimination parameters than items on the cognitive or affective subscales. Two items were flagged for DIF, but the impact of the DIF was negligible. Women scored significantly higher than men on the GAS and its subscales. Participants in the young-old group (60 to 79 years old) scored significantly higher on the cognitive subscale than participants in the old-old group (80 years old and older).Conclusions:Results from the IRT analyses indicated that the GAS and GAS-10 have strong psychometric properties among older adults. We conclude by discussing implications and future research directions.

scholarly journals A Simulation Study to Assess the Effect of the Number of Response Categories on the Power of Ordinal Logistic Regression for Differential Item Functioning Analysis in Rating Scales

2016 ◽  
Vol 2016 ◽  
pp. 1-8 ◽  
Author(s):  
Elahe Allahyari ◽  
Peyman Jafari ◽  
Zahra Bagheri
Keyword(s):  
Logistic Regression ◽  
Sample Size ◽  
Differential Item Functioning ◽  
Rating Scales ◽  
Error Rates ◽  
Ordinal Logistic Regression ◽  
Type I ◽  
Item Functioning ◽  
Quality Of Life Scale ◽  
The Impact

Objective.The present study uses simulated data to find what the optimal number of response categories is to achieve adequate power in ordinal logistic regression (OLR) model for differential item functioning (DIF) analysis in psychometric research.Methods.A hypothetical ten-item quality of life scale with three, four, and five response categories was simulated. The power and type I error rates of OLR model for detecting uniform DIF were investigated under different combinations of ability distribution (θ), sample size, sample size ratio, and the magnitude of uniform DIF across reference and focal groups.Results.Whenθwas distributed identically in the reference and focal groups, increasing the number of response categories from 3 to 5 resulted in an increase of approximately 8% in power of OLR model for detecting uniform DIF. The power of OLR was less than 0.36 when ability distribution in the reference and focal groups was highly skewed to the left and right, respectively.Conclusions.The clearest conclusion from this research is that the minimum number of response categories for DIF analysis using OLR is five. However, the impact of the number of response categories in detecting DIF was lower than might be expected.

Sign in / Sign up

Export Citation Format

Share Document